Bump versions

ci/requirements/py37-min-all-deps.yml

@@ -8,27 +8,27 @@ dependencies:
   # When upgrading python, numpy, or pandas, must also change
   # doc/installing.rst and setup.py.
   - python=3.7
-  - boto3=1.12
+  - boto3=1.13
   - bottleneck=1.3
   - cartopy=0.17
   - cdms2=3.1
   - cfgrib=0.9
-  - cftime=1.0
+  - cftime=1.1
   - coveralls
-  - dask=2.11
-  - distributed=2.11
+  - dask=2.15
+  - distributed=2.15
   - h5netcdf=0.8
   - h5py=2.10
   - hdf5=1.10
   - hypothesis
   - iris=2.4
   - lxml=4.5  # Optional dep of pydap
-  - matplotlib-base=3.1
+  - matplotlib-base=3.2
   - nc-time-axis=1.2
 # netcdf follows a 1.major.minor[.patch] convention (see https://github.com/Unidata/netcdf4-python/issues/1090)
 # bumping the netCDF4 version is currently blocked by #4491
   - netcdf4=1.5.3
-  - numba=0.48
+  - numba=0.49
   - numpy=1.17
   - pandas=1.0
   # - pint  # See py37-min-nep18.yml

doc/whats-new.rst

@@ -95,6 +95,19 @@ New Features
 
 Breaking changes
 ~~~~~~~~~~~~~~~~
+- The minimum versions of some dependencies were changed:
+
+  ============ ====== ====
+  Package      Old    New
+  ============ ====== ====
+  boto3        1.12   1.13
+  cftime       1.0    1.1
+  dask         2.11   2.15
+  distributed  2.11   2.15
+  matplotlib   3.1    3.2
+  numba        0.48   0.49
+  ============ ====== ====
+
 - :py:func:`open_dataset` and :py:func:`open_dataarray` now accept only the first argument
   as positional, all others need to be passed are keyword arguments. This is part of the
   refactor to support external backends (:issue:`4309`, :pull:`4989`).

xarray/core/dask_array_compat.py

@@ -1,6 +1,5 @@
 import warnings
 from distutils.version import LooseVersion
-from typing import Iterable
 
 import numpy as np
 
@@ -58,43 +57,6 @@ def pad(array, pad_width, mode="constant", **kwargs):
     return padded
 
 
-if LooseVersion(dask_version) > LooseVersion("2.9.0"):
-    nanmedian = da.nanmedian
-else:
-
-    def nanmedian(a, axis=None, keepdims=False):
-        """
-        This works by automatically chunking the reduced axes to a single chunk
-        and then calling ``numpy.nanmedian`` function across the remaining dimensions
-        """
-
-        if axis is None:
-            raise NotImplementedError(
-                "The da.nanmedian function only works along an axis.  "
-                "The full algorithm is difficult to do in parallel"
-            )
-
-        if not isinstance(axis, Iterable):
-            axis = (axis,)
-
-        axis = [ax + a.ndim if ax < 0 else ax for ax in axis]
-
-        a = a.rechunk({ax: -1 if ax in axis else "auto" for ax in range(a.ndim)})
-
-        result = da.map_blocks(
-            np.nanmedian,
-            a,
-            axis=axis,
-            keepdims=keepdims,
-            drop_axis=axis if not keepdims else None,
-            chunks=[1 if ax in axis else c for ax, c in enumerate(a.chunks)]
-            if keepdims
-            else None,
-        )
-
-        return result
-
-
 if LooseVersion(dask_version) > LooseVersion("2.30.0"):
     ensure_minimum_chunksize = da.overlap.ensure_minimum_chunksize
 else:

xarray/core/nanops.py

@@ -155,9 +155,7 @@ def nanmedian(a, axis=None, out=None):
     # possibly blow memory
     if axis is not None and len(np.atleast_1d(axis)) == a.ndim:
         axis = None
-    return _dask_or_eager_func(
-        "nanmedian", dask_module=dask_array_compat, eager_module=nputils
-    )(a, axis=axis)
+    return _dask_or_eager_func("nanmedian", eager_module=nputils)(a, axis=axis)
 
 
 def _nanvar_object(value, axis=None, ddof=0, keepdims=False, **kwargs):

xarray/tests/__init__.py

@@ -66,7 +66,6 @@ def LooseVersion(vstring):
 has_pynio, requires_pynio = _importorskip("Nio")
 has_pseudonetcdf, requires_pseudonetcdf = _importorskip("PseudoNetCDF")
 has_cftime, requires_cftime = _importorskip("cftime")
-has_cftime_1_1_0, requires_cftime_1_1_0 = _importorskip("cftime", minversion="1.1.0.0")
 has_cftime_1_4_1, requires_cftime_1_4_1 = _importorskip("cftime", minversion="1.4.1")
 has_dask, requires_dask = _importorskip("dask")
 has_bottleneck, requires_bottleneck = _importorskip("bottleneck")

xarray/tests/test_cftimeindex.py

@@ -16,7 +16,7 @@
 )
 from xarray.tests import assert_array_equal, assert_identical
 
-from . import requires_cftime, requires_cftime_1_1_0
+from . import requires_cftime
 from .test_coding_times import (
     _ALL_CALENDARS,
     _NON_STANDARD_CALENDARS,
@@ -244,7 +244,7 @@ def test_cftimeindex_dayofweek_accessor(index):
     assert_array_equal(result, expected)
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 def test_cftimeindex_days_in_month_accessor(index):
     result = index.days_in_month
     expected = [date.daysinmonth for date in index]
@@ -916,7 +916,7 @@ def test_cftimeindex_calendar_property(calendar, expected):
     assert index.calendar == expected
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 @pytest.mark.parametrize(
     ("calendar", "expected"),
     [
@@ -936,7 +936,7 @@ def test_cftimeindex_calendar_repr(calendar, expected):
     assert "2000-01-01 00:00:00, 2000-01-02 00:00:00" in repr_str
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 @pytest.mark.parametrize("periods", [2, 40])
 def test_cftimeindex_periods_repr(periods):
     """Test that cftimeindex has periods property in repr."""
@@ -945,7 +945,7 @@ def test_cftimeindex_periods_repr(periods):
     assert f" length={periods}" in repr_str
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 @pytest.mark.parametrize("calendar", ["noleap", "360_day", "standard"])
 @pytest.mark.parametrize("freq", ["D", "H"])
 def test_cftimeindex_freq_in_repr(freq, calendar):
@@ -955,7 +955,7 @@ def test_cftimeindex_freq_in_repr(freq, calendar):
     assert f", freq='{freq}'" in repr_str
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 @pytest.mark.parametrize(
     "periods,expected",
     [
@@ -995,7 +995,7 @@ def test_cftimeindex_repr_formatting(periods, expected):
     assert expected == repr(index)
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 @pytest.mark.parametrize("display_width", [40, 80, 100])
 @pytest.mark.parametrize("periods", [2, 3, 4, 100, 101])
 def test_cftimeindex_repr_formatting_width(periods, display_width):
@@ -1013,7 +1013,7 @@ def test_cftimeindex_repr_formatting_width(periods, display_width):
                 assert s[:len_intro_str] == " " * len_intro_str
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 @pytest.mark.parametrize("periods", [22, 50, 100])
 def test_cftimeindex_repr_101_shorter(periods):
     index_101 = xr.cftime_range(start="2000", periods=101)
@@ -1187,7 +1187,7 @@ def test_asi8_distant_date():
     np.testing.assert_array_equal(result, expected)
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 def test_infer_freq_valid_types():
     cf_indx = xr.cftime_range("2000-01-01", periods=3, freq="D")
     assert xr.infer_freq(cf_indx) == "D"
@@ -1202,7 +1202,7 @@ def test_infer_freq_valid_types():
     assert xr.infer_freq(xr.DataArray(pd_td_indx)) == "D"
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 def test_infer_freq_invalid_inputs():
     # Non-datetime DataArray
     with pytest.raises(ValueError, match="must contain datetime-like objects"):
@@ -1231,7 +1231,7 @@ def test_infer_freq_invalid_inputs():
     assert xr.infer_freq(indx[np.array([0, 1, 3])]) is None
 
 
-@requires_cftime_1_1_0
+@requires_cftime
 @pytest.mark.parametrize(
     "freq",
     [